R Notebook
library(tidyverse)Warning: package ‘tidyverse’ was built under R version 4.2.2Registered S3 methods overwritten by 'dbplyr':
method from
print.tbl_lazy
print.tbl_sql
── Attaching packages ──────────────────────────────────────────────────────────── tidyverse 1.3.2 ──✔ ggplot2 3.4.0 ✔ purrr 0.3.4
✔ tibble 3.1.8 ✔ dplyr 1.0.10
✔ tidyr 1.2.1 ✔ stringr 1.4.1
✔ readr 2.1.3 ✔ forcats 0.5.2 Warning: package ‘ggplot2’ was built under R version 4.2.2── Conflicts ─────────────────────────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag() masks stats::lag()library(lubridate)
Attaching package: ‘lubridate’
The following objects are masked from ‘package:base’:
date, intersect, setdiff, unionlibrary(janitor)
Attaching package: ‘janitor’
The following objects are masked from ‘package:stats’:
chisq.test, fisher.testlibrary(caret)Warning: package ‘caret’ was built under R version 4.2.2Loading required package: lattice
Registered S3 method overwritten by 'data.table':
method from
print.data.table
Attaching package: ‘caret’
The following object is masked from ‘package:purrr’:
liftlibrary(broom)Warning: package ‘broom’ was built under R version 4.2.2library(fastDummies)Warning: package ‘fastDummies’ was built under R version 4.2.2library(GGally)Warning: package ‘GGally’ was built under R version 4.2.2Registered S3 method overwritten by 'GGally':
method from
+.gg ggplot2library(ggfortify)Warning: package ‘ggfortify’ was built under R version 4.2.2library(mosaic)Warning: package ‘mosaic’ was built under R version 4.2.2Registered S3 method overwritten by 'mosaic':
method from
fortify.SpatialPolygonsDataFrame ggplot2
The 'mosaic' package masks several functions from core packages in order to add
additional features. The original behavior of these functions should not be affected by this.
Attaching package: ‘mosaic’
The following object is masked from ‘package:Matrix’:
mean
The following object is masked from ‘package:caret’:
dotPlot
The following objects are masked from ‘package:dplyr’:
count, do, tally
The following object is masked from ‘package:purrr’:
cross
The following object is masked from ‘package:ggplot2’:
stat
The following objects are masked from ‘package:stats’:
binom.test, cor, cor.test, cov, fivenum, IQR, median, prop.test, quantile, sd,
t.test, var
The following objects are masked from ‘package:base’:
max, mean, min, prod, range, sample, sumlibrary(mosaicData)
library(modelr)Warning: package ‘modelr’ was built under R version 4.2.2
Attaching package: ‘modelr’
The following object is masked from ‘package:mosaic’:
resample
The following object is masked from ‘package:ggformula’:
na.warn
The following object is masked from ‘package:broom’:
bootstraplibrary(relaimpo)Warning: package ‘relaimpo’ was built under R version 4.2.2Loading required package: MASS
Attaching package: ‘MASS’
The following object is masked from ‘package:dplyr’:
select
Loading required package: boot
Attaching package: ‘boot’
The following object is masked from ‘package:mosaic’:
logit
The following object is masked from ‘package:lattice’:
melanoma
Loading required package: survey
Warning: package ‘survey’ was built under R version 4.2.2Loading required package: grid
Loading required package: survival
Attaching package: ‘survival’
The following object is masked from ‘package:boot’:
aml
The following object is masked from ‘package:caret’:
cluster
Attaching package: ‘survey’
The following object is masked from ‘package:graphics’:
dotchart
Loading required package: mitools
Warning: package ‘mitools’ was built under R version 4.2.2This is the global version of package relaimpo.
If you are a non-US user, a version with the interesting additional metric pmvd is available
from Ulrike Groempings web site at prof.beuth-hochschule.de/groemping.library(tidyverse)
library(glmulti)Warning: package ‘glmulti’ was built under R version 4.2.2Loading required package: rJava
Loading required package: leaps
Warning: package ‘leaps’ was built under R version 4.2.2avocados <- read_csv(here::here("weekend/data/avocado.csv")) %>% clean_names()
New names:Rows: 18249 Columns: 14── Column specification ─────────────────────────────────────────────────────────────────────────────
Delimiter: ","
chr (2): type, region
dbl (11): ...1, AveragePrice, Total Volume, 4046, 4225, 4770, Total Bags, Small Bags, Large Bags...
date (1): Date
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.avocados %>%
distinct(region)
avocados %>%
filter(region == "TotalUS",
date == "2015-12-27")
avocados %>%
filter(region %in% c("Midsouth", "Northeast", "Plains", "SouthCentral", "Southeast", "West", "GreatLakes", "California"),
date == "2015-12-27") %>%
group_by(type) %>%
summarise(total_volume = sum(total_volume),
x4046 = sum(x4046),
total_bags = sum(total_bags))
avocados %>%
filter(!region %in% c("TotalUS", "Midsouth", "Northeast", "Plains", "SouthCentral", "Southeast", "West", "GreatLakes", "California"),
date == "2015-12-27") %>%
group_by(type) %>%
summarise(total_volume = sum(total_volume),
x4046 = sum(x4046),
total_bags = sum(total_bags))
# Missouth, Northeast, Plains, SouthCentral, Southeast, West, GreatLakes, California
# These are the regions that make up the total US. Above code shows that the total numbers of the categories are the same
# I will filter the data so that it only uses the lines for the above regions to avoid double counting.
# no need to worry about the smaller areas as these are contained within the bigger regions
# also i dont think it makes sense to look at regions that small
# also by the looks of it the remaining cities arent an exhaustive list
# might be good to extract month from date as a categorical
# proportions of 4046, 4225, 4770 may be better than absolutes, also props of bag size may be better than absolute numbers avocados %>%
filter(region %in% c("Midsouth",
"Northeast",
"Plains",
"SouthCentral",
"Southeast",
"West",
"GreatLakes",
"California")) %>%
mutate(a = round(small_bags + large_bags + x_large_bags, 2) - round(total_bags, 2)) %>%
arrange(desc(a))
avocados
alias(lm(average_price ~ ., data = avocados))Model :
average_price ~ x1 + date + total_volume + x4046 + x4225 + x4770 +
total_bags + small_bags + large_bags + x_large_bags + type +
year + region# for some reason the bag sizes dont properly add up to the total_bags
# without the round they are different by like 0.000000000001
# with the round there are a couple at either end out by +-1
# this means they dont show up in alias - I'm going to manual remove x larges bags as this can be derived from the other 3
# x4046, x4225 and x 4770 dont add up to the total volume so can keep them allavocados <- avocados %>% filter(region %in% c("Midsouth",
"Northeast",
"Plains",
"SouthCentral",
"Southeast",
"West",
"GreatLakes",
"California")) %>%
mutate(month = as.factor(month(date)),
across(where(is.character), as.factor),
year = as.factor(year)) %>%
dplyr::select(-x1, -date, -x_large_bags)
avocados %>%
dplyr::select(average_price, is.factor) %>%
ggpairs() Warning: Use of bare predicate functions was deprecated in tidyselect 1.1.0.
Please use wrap predicates in `where()` instead.
# Was:
data %>% select(is.factor)
# Now:
data %>% select(where(is.factor))
# month, type and region all quite strong
avocados %>%
dplyr::select(is.numeric) %>%
ggpairs() Warning: Use of bare predicate functions was deprecated in tidyselect 1.1.0.
Please use wrap predicates in `where()` instead.
# Was:
data %>% select(is.numeric)
# Now:
data %>% select(where(is.numeric))
# total volume and total bags strongest
avocados_prop <- avocados %>%
mutate(across(small_bags:large_bags, ~ .x / total_bags))
avocados_prop %>%
dplyr::select(is.numeric) %>%
ggpairs() 
#props don really help
n_data <- nrow(avocados_prop)
test_index <- sample(1:n_data, size = n_data * 0.2)
test <- avocados_prop %>%
slice(test_index)
train <- avocados_prop %>%
slice(-test_index)lm1 <- lm(average_price ~ total_volume, train)
summary(lm1)
Call:
lm(formula = average_price ~ total_volume, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.73542 -0.17104 -0.01301 0.14873 0.98879
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.599e+00 7.101e-03 225.20 <2e-16 ***
total_volume -1.061e-07 2.193e-09 -48.41 <2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.2452 on 2162 degrees of freedom
Multiple R-squared: 0.5201, Adjusted R-squared: 0.5199
F-statistic: 2343 on 1 and 2162 DF, p-value: < 2.2e-16plotModel(lm1)
plot(lm1)
# diagnostocs are mad i think this is because of the organic type
avocados_resid <- train %>%
add_residuals(lm1) %>%
dplyr::select(-average_price, -total_volume) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(resid, is.factor) %>%
ggpairs() 
# month, type and region all quite strong
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs()
train %>%
ggplot(aes(x = x4046, y = average_price, colour = type)) +
geom_point() +
geom_smooth(method = "lm", se = FALSE)
# filter by type to check if the are correlated with different things
avocados %>%
filter(type == "conventional") %>%
dplyr::select(average_price, is.factor) %>%
ggpairs() 
# month, type and region all quite strong
avocados %>%
filter(type == "conventional") %>%
dplyr::select(is.numeric) %>%
ggpairs() 
# total volume and total bags strongest
avocados_prop %>%
filter(type == "conventional") %>%
dplyr::select(is.numeric) %>%
ggpairs()
avocados %>%
filter(type == "organic") %>%
dplyr::select(average_price, is.factor) %>%
ggpairs() 
# month, type and region all quite strong
avocados %>%
filter(type == "organic") %>%
dplyr::select(is.numeric) %>%
ggpairs() 
# total volume and total bags strongest
avocados_prop %>%
filter(type == "organic") %>%
dplyr::select(is.numeric) %>%
ggpairs() 
#props don really help # going to do type first
lm2 <- lm(average_price ~ type, train)
summary(lm2)
Call:
lm(formula = average_price ~ type, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.78247 -0.17247 -0.01781 0.15702 0.96753
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.123148 0.007792 144.14 <2e-16 ***
typeorganic 0.489327 0.010994 44.51 <2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.2557 on 2162 degrees of freedom
Multiple R-squared: 0.4781, Adjusted R-squared: 0.4779
F-statistic: 1981 on 1 and 2162 DF, p-value: < 2.2e-16plot(lm2)
NA
NAavocados_resid <- train %>%
add_residuals(lm2) %>%
dplyr::select(-average_price, -type) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(resid, is.factor) %>%
ggpairs() 
# month, type and region all quite strong
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs() 
lm3 <- lm(average_price ~ type + region, train)
summary(lm3)
Call:
lm(formula = average_price ~ type + region, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.74699 -0.13834 -0.02354 0.11732 0.99309
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.159535 0.014116 82.142 < 2e-16 ***
typeorganic 0.489463 0.009401 52.064 < 2e-16 ***
regionGreatLakes -0.072013 0.018815 -3.827 0.000133 ***
regionMidsouth -0.001411 0.018798 -0.075 0.940167
regionNortheast 0.194000 0.018869 10.281 < 2e-16 ***
regionPlains 0.031770 0.018763 1.693 0.090557 .
regionSouthCentral -0.302675 0.018646 -16.233 < 2e-16 ***
regionSoutheast -0.007822 0.018835 -0.415 0.677976
regionWest -0.122087 0.018815 -6.489 1.07e-10 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.2186 on 2155 degrees of freedom
Multiple R-squared: 0.6198, Adjusted R-squared: 0.6184
F-statistic: 439.2 on 8 and 2155 DF, p-value: < 2.2e-16plot(lm3)
anova(lm2, lm3)Analysis of Variance Table
Model 1: average_price ~ type
Model 2: average_price ~ type + region
Res.Df RSS Df Sum of Sq F Pr(>F)
1 2162 141.38
2 2155 102.99 7 38.391 114.76 < 2.2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1avocados_resid <- train %>%
add_residuals(lm3) %>%
dplyr::select(-average_price, -type, -region) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(resid, is.factor) %>%
ggpairs() 
# month, type and region all quite strong
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs() 
lm4 <- lm(average_price ~ type + region + month, train)
summary(lm4)
Call:
lm(formula = average_price ~ type + region + month, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.65276 -0.11864 -0.00312 0.10866 0.82500
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.062100 0.017046 62.308 < 2e-16 ***
typeorganic 0.491225 0.008080 60.792 < 2e-16 ***
regionGreatLakes -0.068198 0.016167 -4.218 2.56e-05 ***
regionMidsouth 0.006574 0.016154 0.407 0.68408
regionNortheast 0.196708 0.016215 12.131 < 2e-16 ***
regionPlains 0.034712 0.016121 2.153 0.03141 *
regionSouthCentral -0.303184 0.016021 -18.924 < 2e-16 ***
regionSoutheast -0.004919 0.016186 -0.304 0.76121
regionWest -0.125033 0.016174 -7.731 1.63e-14 ***
month2 -0.055798 0.018249 -3.058 0.00226 **
month3 -0.002364 0.018069 -0.131 0.89591
month4 0.043316 0.019043 2.275 0.02302 *
month5 0.026449 0.018635 1.419 0.15596
month6 0.087948 0.019977 4.403 1.12e-05 ***
month7 0.150393 0.018715 8.036 1.52e-15 ***
month8 0.206713 0.019271 10.726 < 2e-16 ***
month9 0.283468 0.019550 14.500 < 2e-16 ***
month10 0.304755 0.018833 16.182 < 2e-16 ***
month11 0.152789 0.019108 7.996 2.08e-15 ***
month12 0.024587 0.019042 1.291 0.19678
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.1878 on 2144 degrees of freedom
Multiple R-squared: 0.7209, Adjusted R-squared: 0.7185
F-statistic: 291.5 on 19 and 2144 DF, p-value: < 2.2e-16plot(lm4)
anova(lm3, lm4)Analysis of Variance Table
Model 1: average_price ~ type + region
Model 2: average_price ~ type + region + month
Res.Df RSS Df Sum of Sq F Pr(>F)
1 2155 102.990
2 2144 75.605 11 27.385 70.598 < 2.2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1# log helps diagnostics
lm4 <- lm(log(average_price) ~ type + region + month, train)
summary(lm4)
Call:
lm(formula = log(average_price) ~ type + region + month, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.57536 -0.08972 -0.00062 0.08946 0.44429
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.043982 0.012402 3.546 0.000399 ***
typeorganic 0.368973 0.005879 62.760 < 2e-16 ***
regionGreatLakes -0.027132 0.011762 -2.307 0.021166 *
regionMidsouth 0.027394 0.011753 2.331 0.019859 *
regionNortheast 0.152288 0.011797 12.909 < 2e-16 ***
regionPlains 0.036724 0.011729 3.131 0.001765 **
regionSouthCentral -0.236958 0.011657 -20.328 < 2e-16 ***
regionSoutheast 0.009883 0.011776 0.839 0.401447
regionWest -0.096932 0.011767 -8.237 3.03e-16 ***
month2 -0.051697 0.013277 -3.894 0.000102 ***
month3 0.003204 0.013147 0.244 0.807467
month4 0.033313 0.013855 2.404 0.016283 *
month5 0.012766 0.013558 0.942 0.346503
month6 0.066478 0.014535 4.574 5.06e-06 ***
month7 0.116553 0.013616 8.560 < 2e-16 ***
month8 0.150148 0.014021 10.709 < 2e-16 ***
month9 0.197583 0.014224 13.891 < 2e-16 ***
month10 0.221953 0.013702 16.198 < 2e-16 ***
month11 0.115575 0.013903 8.313 < 2e-16 ***
month12 0.020065 0.013854 1.448 0.147681
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.1366 on 2144 degrees of freedom
Multiple R-squared: 0.7365, Adjusted R-squared: 0.7342
F-statistic: 315.5 on 19 and 2144 DF, p-value: < 2.2e-16plot(lm4)
NA
avocados_resid <- train %>%
add_residuals(lm4) %>%
dplyr::select(-average_price, -type, -region, -month) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(resid, is.factor) %>%
ggpairs() 
# month, type and region all quite strong
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs() 
lm5 <- lm(log(average_price) ~ type + region + month + year, train)
summary(lm5)
Call:
lm(formula = log(average_price) ~ type + region + month + year,
data = train)
Residuals:
Min 1Q Median 3Q Max
-0.62182 -0.07801 0.00214 0.08776 0.38374
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.010441 0.012551 0.832 0.405604
typeorganic 0.368094 0.005469 67.302 < 2e-16 ***
regionGreatLakes -0.027873 0.010941 -2.548 0.010913 *
regionMidsouth 0.026756 0.010933 2.447 0.014476 *
regionNortheast 0.151809 0.010973 13.834 < 2e-16 ***
regionPlains 0.038806 0.010912 3.556 0.000384 ***
regionSouthCentral -0.235747 0.010844 -21.739 < 2e-16 ***
regionSoutheast 0.011921 0.010956 1.088 0.276685
regionWest -0.097710 0.010946 -8.927 < 2e-16 ***
month2 -0.056071 0.012362 -4.536 6.05e-06 ***
month3 -0.001403 0.012255 -0.114 0.908878
month4 0.043917 0.013118 3.348 0.000829 ***
month5 0.033542 0.012852 2.610 0.009122 **
month6 0.081417 0.013738 5.926 3.60e-09 ***
month7 0.129660 0.012899 10.052 < 2e-16 ***
month8 0.168997 0.013277 12.728 < 2e-16 ***
month9 0.210681 0.013457 15.656 < 2e-16 ***
month10 0.233360 0.012978 17.982 < 2e-16 ***
month11 0.130525 0.013170 9.911 < 2e-16 ***
month12 0.031657 0.013118 2.413 0.015895 *
year2016 -0.029911 0.006979 -4.286 1.90e-05 ***
year2017 0.086228 0.006928 12.446 < 2e-16 ***
year2018 0.090947 0.012529 7.259 5.43e-13 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.1271 on 2141 degrees of freedom
Multiple R-squared: 0.7724, Adjusted R-squared: 0.7701
F-statistic: 330.3 on 22 and 2141 DF, p-value: < 2.2e-16plot(lm5)
anova(lm4, lm5)Analysis of Variance Table
Model 1: log(average_price) ~ type + region + month
Model 2: log(average_price) ~ type + region + month + year
Res.Df RSS Df Sum of Sq F Pr(>F)
1 2144 40.022
2 2141 34.575 3 5.4472 112.44 < 2.2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1avocados_resid <- train %>%
add_residuals(lm5) %>%
dplyr::select(-average_price, -type, -region, -month, -year) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(resid, is.factor) %>%
ggpairs() 
# month, type and region all quite strong
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs() 
lm6 <- lm(log(average_price) ~ type + region + month + year + x4046, train)
summary(lm6)
Call:
lm(formula = log(average_price) ~ type + region + month + year +
x4046, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.59343 -0.07397 0.00511 0.08087 0.37116
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.417e-01 1.489e-02 9.520 < 2e-16 ***
typeorganic 2.667e-01 8.608e-03 30.982 < 2e-16 ***
regionGreatLakes -8.766e-02 1.118e-02 -7.842 6.95e-15 ***
regionMidsouth -3.016e-02 1.110e-02 -2.716 0.006656 **
regionNortheast 7.905e-02 1.155e-02 6.843 1.01e-11 ***
regionPlains -1.125e-02 1.093e-02 -1.029 0.303418
regionSouthCentral -2.070e-01 1.051e-02 -19.695 < 2e-16 ***
regionSoutheast 2.699e-03 1.046e-02 0.258 0.796351
regionWest -9.815e-02 1.043e-02 -9.413 < 2e-16 ***
month2 -4.407e-02 1.180e-02 -3.733 0.000194 ***
month3 -1.487e-05 1.168e-02 -0.001 0.998984
month4 4.874e-02 1.250e-02 3.899 9.97e-05 ***
month5 4.678e-02 1.228e-02 3.810 0.000143 ***
month6 8.599e-02 1.309e-02 6.568 6.38e-11 ***
month7 1.291e-01 1.229e-02 10.508 < 2e-16 ***
month8 1.694e-01 1.265e-02 13.389 < 2e-16 ***
month9 2.014e-01 1.283e-02 15.689 < 2e-16 ***
month10 2.181e-01 1.241e-02 17.578 < 2e-16 ***
month11 1.191e-01 1.257e-02 9.477 < 2e-16 ***
month12 2.312e-02 1.251e-02 1.848 0.064770 .
year2016 -3.433e-02 6.656e-03 -5.159 2.72e-07 ***
year2017 8.665e-02 6.600e-03 13.129 < 2e-16 ***
year2018 9.686e-02 1.194e-02 8.111 8.38e-16 ***
x4046 -6.759e-08 4.567e-09 -14.800 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.1211 on 2140 degrees of freedom
Multiple R-squared: 0.7935, Adjusted R-squared: 0.7913
F-statistic: 357.6 on 23 and 2140 DF, p-value: < 2.2e-16plot(lm6)
NA
avocados_resid <- train %>%
add_residuals(lm6) %>%
dplyr::select(-average_price, -type, -region, -month, -year, -x4046) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs() 
lm7 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume, train)
summary(lm7)
Call:
lm(formula = log(average_price) ~ type + region + month + year +
x4046 + total_volume, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.59290 -0.06734 0.00637 0.07724 0.35848
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.658e-01 1.803e-02 14.741 < 2e-16 ***
typeorganic 1.591e-01 1.254e-02 12.682 < 2e-16 ***
regionGreatLakes -1.001e-01 1.090e-02 -9.183 < 2e-16 ***
regionMidsouth -5.575e-02 1.100e-02 -5.067 4.39e-07 ***
regionNortheast 9.375e-02 1.129e-02 8.307 < 2e-16 ***
regionPlains -6.756e-02 1.168e-02 -5.782 8.48e-09 ***
regionSouthCentral -2.321e-01 1.043e-02 -22.245 < 2e-16 ***
regionSoutheast -4.491e-02 1.096e-02 -4.098 4.33e-05 ***
regionWest -9.192e-02 1.014e-02 -9.069 < 2e-16 ***
month2 -3.620e-02 1.148e-02 -3.154 0.001635 **
month3 -1.842e-03 1.133e-02 -0.163 0.870920
month4 4.809e-02 1.213e-02 3.963 7.65e-05 ***
month5 5.242e-02 1.193e-02 4.396 1.16e-05 ***
month6 8.722e-02 1.271e-02 6.863 8.78e-12 ***
month7 1.292e-01 1.193e-02 10.829 < 2e-16 ***
month8 1.620e-01 1.229e-02 13.174 < 2e-16 ***
month9 1.942e-01 1.247e-02 15.567 < 2e-16 ***
month10 2.062e-01 1.209e-02 17.061 < 2e-16 ***
month11 1.079e-01 1.224e-02 8.817 < 2e-16 ***
month12 1.886e-02 1.215e-02 1.552 0.120730
year2016 -2.258e-02 6.540e-03 -3.453 0.000565 ***
year2017 9.830e-02 6.486e-03 15.156 < 2e-16 ***
year2018 1.182e-01 1.174e-02 10.071 < 2e-16 ***
x4046 -1.456e-08 6.394e-09 -2.277 0.022905 *
total_volume -4.551e-08 3.955e-09 -11.507 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.1175 on 2139 degrees of freedom
Multiple R-squared: 0.8056, Adjusted R-squared: 0.8034
F-statistic: 369.3 on 24 and 2139 DF, p-value: < 2.2e-16plot(lm7)
avocados_resid <- train %>%
add_residuals(lm7) %>%
dplyr::select(-average_price, -type, -region, -month, -year, -x4046, -total_volume) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs() 
lm8 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags, train)
lm8b <- lm(log(average_price)~1+type+year+region+month+total_volume+total_bags+small_bags+large_bags, train)
summary(lm8)
Call:
lm(formula = log(average_price) ~ type + region + month + year +
x4046 + total_volume + large_bags, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.53786 -0.06832 0.00604 0.07984 0.38228
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.486e-01 1.795e-02 13.852 < 2e-16 ***
typeorganic 1.811e-01 1.273e-02 14.223 < 2e-16 ***
regionGreatLakes -5.434e-02 1.239e-02 -4.384 1.22e-05 ***
regionMidsouth -3.020e-02 1.139e-02 -2.651 0.008097 **
regionNortheast 1.069e-01 1.128e-02 9.475 < 2e-16 ***
regionPlains -4.086e-02 1.208e-02 -3.382 0.000732 ***
regionSouthCentral -2.233e-01 1.037e-02 -21.537 < 2e-16 ***
regionSoutheast -9.029e-03 1.184e-02 -0.762 0.445942
regionWest -3.452e-02 1.262e-02 -2.735 0.006298 **
month2 -3.695e-02 1.133e-02 -3.260 0.001133 **
month3 -2.796e-03 1.119e-02 -0.250 0.802777
month4 4.876e-02 1.198e-02 4.069 4.88e-05 ***
month5 5.270e-02 1.178e-02 4.475 8.04e-06 ***
month6 8.732e-02 1.255e-02 6.959 4.55e-12 ***
month7 1.282e-01 1.178e-02 10.884 < 2e-16 ***
month8 1.566e-01 1.216e-02 12.874 < 2e-16 ***
month9 1.927e-01 1.232e-02 15.643 < 2e-16 ***
month10 2.065e-01 1.193e-02 17.300 < 2e-16 ***
month11 1.094e-01 1.209e-02 9.054 < 2e-16 ***
month12 2.001e-02 1.200e-02 1.668 0.095498 .
year2016 -1.371e-02 6.567e-03 -2.088 0.036908 *
year2017 1.033e-01 6.439e-03 16.040 < 2e-16 ***
year2018 1.171e-01 1.159e-02 10.097 < 2e-16 ***
x4046 -9.166e-09 6.355e-09 -1.442 0.149394
total_volume -4.477e-08 3.907e-09 -11.459 < 2e-16 ***
large_bags -1.268e-01 1.700e-02 -7.459 1.26e-13 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.116 on 2138 degrees of freedom
Multiple R-squared: 0.8105, Adjusted R-squared: 0.8083
F-statistic: 365.8 on 25 and 2138 DF, p-value: < 2.2e-16plot(lm8)
summary(lm8b)
Call:
lm(formula = log(average_price) ~ 1 + type + year + region +
month + total_volume + total_bags + small_bags + large_bags,
data = train)
Residuals:
Min 1Q Median 3Q Max
-0.53155 -0.06898 0.00640 0.07745 0.38245
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.087e+00 1.895e-01 5.737 1.10e-08 ***
typeorganic 1.895e-01 1.297e-02 14.612 < 2e-16 ***
year2016 -2.484e-02 6.945e-03 -3.577 0.000356 ***
year2017 8.886e-02 7.120e-03 12.480 < 2e-16 ***
year2018 9.844e-02 1.225e-02 8.039 1.49e-15 ***
regionGreatLakes -6.579e-02 1.233e-02 -5.334 1.06e-07 ***
regionMidsouth -2.953e-02 1.142e-02 -2.585 0.009798 **
regionNortheast 1.159e-01 1.049e-02 11.044 < 2e-16 ***
regionPlains -4.236e-02 1.205e-02 -3.516 0.000447 ***
regionSouthCentral -2.226e-01 9.986e-03 -22.290 < 2e-16 ***
regionSoutheast -1.606e-02 1.175e-02 -1.367 0.171841
regionWest -4.054e-02 1.296e-02 -3.129 0.001778 **
month2 -3.789e-02 1.126e-02 -3.366 0.000777 ***
month3 -6.334e-03 1.113e-02 -0.569 0.569436
month4 4.466e-02 1.192e-02 3.747 0.000184 ***
month5 4.934e-02 1.171e-02 4.214 2.62e-05 ***
month6 7.738e-02 1.259e-02 6.145 9.50e-10 ***
month7 1.165e-01 1.191e-02 9.783 < 2e-16 ***
month8 1.506e-01 1.211e-02 12.434 < 2e-16 ***
month9 1.895e-01 1.225e-02 15.475 < 2e-16 ***
month10 2.035e-01 1.187e-02 17.136 < 2e-16 ***
month11 1.053e-01 1.204e-02 8.747 < 2e-16 ***
month12 1.358e-02 1.198e-02 1.134 0.257117
total_volume -5.832e-08 3.857e-09 -15.120 < 2e-16 ***
total_bags 3.612e-08 9.623e-09 3.754 0.000179 ***
small_bags -8.341e-01 1.915e-01 -4.356 1.39e-05 ***
large_bags -9.535e-01 1.908e-01 -4.996 6.32e-07 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.1153 on 2137 degrees of freedom
Multiple R-squared: 0.8131, Adjusted R-squared: 0.8109
F-statistic: 357.7 on 26 and 2137 DF, p-value: < 2.2e-16plot(lm8b)
avocados_resid <- train %>%
add_residuals(lm8) %>%
dplyr::select(-average_price, -type, -region, -month, -year, -x4046, -total_volume, -large_bags) %>%
dplyr::select(resid, everything())
avocados_resid %>%
dplyr::select(is.numeric) %>%
ggpairs() 
lm9 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags, train)
summary(lm9)
Call:
lm(formula = log(average_price) ~ type + region + month + year +
x4046 + total_volume + large_bags, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.53786 -0.06832 0.00604 0.07984 0.38228
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.486e-01 1.795e-02 13.852 < 2e-16 ***
typeorganic 1.811e-01 1.273e-02 14.223 < 2e-16 ***
regionGreatLakes -5.434e-02 1.239e-02 -4.384 1.22e-05 ***
regionMidsouth -3.020e-02 1.139e-02 -2.651 0.008097 **
regionNortheast 1.069e-01 1.128e-02 9.475 < 2e-16 ***
regionPlains -4.086e-02 1.208e-02 -3.382 0.000732 ***
regionSouthCentral -2.233e-01 1.037e-02 -21.537 < 2e-16 ***
regionSoutheast -9.029e-03 1.184e-02 -0.762 0.445942
regionWest -3.452e-02 1.262e-02 -2.735 0.006298 **
month2 -3.695e-02 1.133e-02 -3.260 0.001133 **
month3 -2.796e-03 1.119e-02 -0.250 0.802777
month4 4.876e-02 1.198e-02 4.069 4.88e-05 ***
month5 5.270e-02 1.178e-02 4.475 8.04e-06 ***
month6 8.732e-02 1.255e-02 6.959 4.55e-12 ***
month7 1.282e-01 1.178e-02 10.884 < 2e-16 ***
month8 1.566e-01 1.216e-02 12.874 < 2e-16 ***
month9 1.927e-01 1.232e-02 15.643 < 2e-16 ***
month10 2.065e-01 1.193e-02 17.300 < 2e-16 ***
month11 1.094e-01 1.209e-02 9.054 < 2e-16 ***
month12 2.001e-02 1.200e-02 1.668 0.095498 .
year2016 -1.371e-02 6.567e-03 -2.088 0.036908 *
year2017 1.033e-01 6.439e-03 16.040 < 2e-16 ***
year2018 1.171e-01 1.159e-02 10.097 < 2e-16 ***
x4046 -9.166e-09 6.355e-09 -1.442 0.149394
total_volume -4.477e-08 3.907e-09 -11.459 < 2e-16 ***
large_bags -1.268e-01 1.700e-02 -7.459 1.26e-13 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.116 on 2138 degrees of freedom
Multiple R-squared: 0.8105, Adjusted R-squared: 0.8083
F-statistic: 365.8 on 25 and 2138 DF, p-value: < 2.2e-16plot(lm9)
lm10 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume, train)
summary(lm10)
Call:
lm(formula = log(average_price) ~ type + region + month + year +
x4046 + total_volume + large_bags + type:total_volume, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.36543 -0.06091 0.00252 0.06397 0.68320
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 3.783e-01 1.609e-02 23.509 < 2e-16 ***
typeorganic 2.288e-01 1.106e-02 20.689 < 2e-16 ***
regionGreatLakes -1.098e-01 1.082e-02 -10.146 < 2e-16 ***
regionMidsouth -1.102e-01 1.019e-02 -10.812 < 2e-16 ***
regionNortheast 1.014e-01 9.681e-03 10.479 < 2e-16 ***
regionPlains -1.938e-01 1.174e-02 -16.504 < 2e-16 ***
regionSouthCentral -2.785e-01 9.115e-03 -30.557 < 2e-16 ***
regionSoutheast -1.407e-01 1.122e-02 -12.547 < 2e-16 ***
regionWest -3.122e-03 1.089e-02 -0.287 0.774343
month2 -1.405e-02 9.758e-03 -1.440 0.150073
month3 2.592e-02 9.657e-03 2.684 0.007335 **
month4 7.629e-02 1.033e-02 7.388 2.13e-13 ***
month5 8.944e-02 1.019e-02 8.778 < 2e-16 ***
month6 1.113e-01 1.080e-02 10.302 < 2e-16 ***
month7 1.439e-01 1.012e-02 14.220 < 2e-16 ***
month8 1.640e-01 1.044e-02 15.716 < 2e-16 ***
month9 1.920e-01 1.057e-02 18.172 < 2e-16 ***
month10 1.955e-01 1.025e-02 19.078 < 2e-16 ***
month11 9.824e-02 1.038e-02 9.466 < 2e-16 ***
month12 2.400e-02 1.029e-02 2.331 0.019820 *
year2016 2.060e-02 5.768e-03 3.571 0.000363 ***
year2017 1.636e-01 5.937e-03 27.559 < 2e-16 ***
year2018 2.257e-01 1.069e-02 21.113 < 2e-16 ***
x4046 2.991e-08 5.631e-09 5.311 1.20e-07 ***
total_volume -9.027e-08 3.732e-09 -24.189 < 2e-16 ***
large_bags -6.960e-02 1.473e-02 -4.726 2.44e-06 ***
typeorganic:total_volume -1.514e-06 5.461e-08 -27.720 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.09954 on 2137 degrees of freedom
Multiple R-squared: 0.8606, Adjusted R-squared: 0.8589
F-statistic: 507.5 on 26 and 2137 DF, p-value: < 2.2e-16plot(lm10)
lm11 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume + type:region, train)
lm12 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume + type:region + type:year + region:total_volume, train)
summary(lm11)
Call:
lm(formula = log(average_price) ~ type + region + month + year +
x4046 + total_volume + large_bags + type:total_volume + type:region,
data = train)
Residuals:
Min 1Q Median 3Q Max
-0.31319 -0.05543 0.00383 0.05520 0.73756
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 5.338e-01 2.367e-02 22.546 < 2e-16 ***
typeorganic 7.485e-02 2.454e-02 3.050 0.00232 **
regionGreatLakes -1.521e-01 1.697e-02 -8.962 < 2e-16 ***
regionMidsouth -1.975e-01 1.675e-02 -11.789 < 2e-16 ***
regionNortheast 3.109e-02 1.913e-02 1.625 0.10423
regionPlains -3.585e-01 1.848e-02 -19.405 < 2e-16 ***
regionSouthCentral -2.141e-01 1.352e-02 -15.836 < 2e-16 ***
regionSoutheast -1.235e-01 1.551e-02 -7.962 2.73e-15 ***
regionWest -2.386e-02 1.191e-02 -2.004 0.04522 *
month2 -8.440e-03 8.785e-03 -0.961 0.33682
month3 2.716e-02 8.595e-03 3.161 0.00160 **
month4 7.934e-02 9.420e-03 8.423 < 2e-16 ***
month5 9.274e-02 9.491e-03 9.771 < 2e-16 ***
month6 1.093e-01 9.581e-03 11.408 < 2e-16 ***
month7 1.409e-01 9.426e-03 14.950 < 2e-16 ***
month8 1.524e-01 9.400e-03 16.211 < 2e-16 ***
month9 1.735e-01 9.690e-03 17.899 < 2e-16 ***
month10 1.677e-01 9.480e-03 17.684 < 2e-16 ***
month11 9.017e-02 9.611e-03 9.382 < 2e-16 ***
month12 8.005e-03 9.404e-03 0.851 0.39475
year2016 3.749e-02 5.388e-03 6.959 4.56e-12 ***
year2017 1.765e-01 5.507e-03 32.056 < 2e-16 ***
year2018 2.355e-01 9.641e-03 24.426 < 2e-16 ***
x4046 3.031e-09 9.240e-09 0.328 0.74296
total_volume -1.029e-07 4.703e-09 -21.882 < 2e-16 ***
large_bags -2.036e-01 1.540e-02 -13.223 < 2e-16 ***
typeorganic:total_volume -1.574e-06 5.197e-08 -30.280 < 2e-16 ***
typeorganic:regionGreatLakes 9.036e-02 2.020e-02 4.474 8.09e-06 ***
typeorganic:regionMidsouth 1.277e-01 2.008e-02 6.357 2.51e-10 ***
typeorganic:regionNortheast 7.221e-02 2.194e-02 3.291 0.00101 **
typeorganic:regionPlains 2.662e-01 2.190e-02 12.155 < 2e-16 ***
typeorganic:regionSouthCentral -8.674e-02 1.753e-02 -4.947 8.12e-07 ***
typeorganic:regionSoutheast -1.484e-02 1.879e-02 -0.790 0.42979
typeorganic:regionWest 1.707e-01 1.667e-02 10.242 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.09041 on 2130 degrees of freedom
Multiple R-squared: 0.8822, Adjusted R-squared: 0.8804
F-statistic: 483.4 on 33 and 2130 DF, p-value: < 2.2e-16plot(lm11)
glance(lm11)glance(lm12)train_rmse <- train %>%
add_residuals(lm12) %>%
mutate(sq_resid = resid^2) %>%
summarise(mse = mean(sq_resid),
rmse = mse^0.5) %>%
pull(rmse)
train_rmse[1] 0.08437852predictions_test <- test %>%
add_predictions(lm12) %>%
add_residuals(lm12) %>%
dplyr::select(average_price, pred, resid)
test_rsme <- predictions_test %>%
mutate(sq_resid = resid^2) %>%
summarise(mse = mean(sq_resid),
rmse = mse^0.5) %>%
pull(rmse)
test_rsme / train_rmse[1] 0.9530149cv_10_fold <- trainControl(
method = "cv",
number = 10,
savePredictions = TRUE
)
model2 <- train(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume,
avocados_prop,
trControl = cv_10_fold,
method = "lm")
model2$resample %>%
summarise(av_r2 = mean(Rsquared),
av_rmse = mean(RMSE))cv_10_fold <- trainControl(
method = "cv",
number = 10,
savePredictions = TRUE
)
model2 <- train(log(average_price) ~ type + region + month + year + x4046+ total_volume + large_bags + type:total_volume + type:region + type:year + region:total_volume,
avocados_prop,
trControl = cv_10_fold,
method = "lm")
model2$resample %>%
summarise(av_r2 = mean(Rsquared),
av_rmse = mean(RMSE))calc.relimp(lm9, type = "lmg", rela = TRUE)Response variable: log(average_price)
Total response variance: 0.07023302
Analysis based on 2164 observations
25 Regressors:
Some regressors combined in groups:
Group region : regionGreatLakes regionMidsouth regionNortheast regionPlains regionSouthCentral regionSoutheast regionWest
Group month : month2 month3 month4 month5 month6 month7 month8 month9 month10 month11 month12
Group year : year2016 year2017 year2018
Relative importance of 7 (groups of) regressors assessed:
region month year type x4046 total_volume large_bags
Proportion of variance explained by model: 81.05%
Metrics are normalized to sum to 100% (rela=TRUE).
Relative importance metrics:
lmg
region 0.134641265
month 0.095970421
year 0.048764397
type 0.226128036
x4046 0.226597172
total_volume 0.261021543
large_bags 0.006877166
Average coefficients for different model sizes:
1group 2groups 3groups 4groups 5groups
type 3.678353e-01 2.947586e-01 2.436647e-01 2.119444e-01 1.952180e-01
regionGreatLakes -2.975501e-02 -8.130292e-02 -1.008357e-01 -9.960861e-02 -8.664705e-02
regionMidsouth 2.099359e-02 -2.820968e-02 -5.283919e-02 -5.923284e-02 -5.349555e-02
regionNortheast 1.442123e-01 9.997769e-02 7.877882e-02 7.564505e-02 8.432104e-02
regionPlains 3.675087e-02 -1.894309e-02 -5.004991e-02 -6.243166e-02 -6.147231e-02
regionSouthCentral -2.418414e-01 -2.290649e-01 -2.213823e-01 -2.190153e-01 -2.205260e-01
regionSoutheast -2.309713e-03 -2.357879e-02 -3.084214e-02 -3.041651e-02 -2.584134e-02
regionWest -9.498219e-02 -1.033868e-01 -9.278081e-02 -7.559419e-02 -5.887311e-02
month2 -5.322182e-02 -4.215024e-02 -3.489426e-02 -3.177629e-02 -3.219225e-02
month3 2.336431e-02 1.566266e-02 1.008000e-02 5.845023e-03 2.510236e-03
month4 4.610019e-02 4.461026e-02 4.471396e-02 4.574120e-02 4.695995e-02
month5 9.289308e-03 2.413591e-02 3.603635e-02 4.431730e-02 4.895355e-02
month6 8.095504e-02 7.946350e-02 8.010731e-02 8.194235e-02 8.398654e-02
month7 1.096422e-01 1.128397e-01 1.159579e-01 1.192298e-01 1.224712e-01
month8 1.477032e-01 1.483213e-01 1.495314e-01 1.511501e-01 1.530281e-01
month9 1.953311e-01 1.895520e-01 1.861248e-01 1.854024e-01 1.870272e-01
month10 2.185677e-01 2.077589e-01 2.008506e-01 1.984080e-01 1.997708e-01
month11 1.220613e-01 1.095958e-01 1.025676e-01 1.005300e-01 1.022746e-01
month12 2.273399e-02 1.494875e-02 1.076900e-02 1.031215e-02 1.268258e-02
year2016 -3.709716e-02 -3.284356e-02 -2.870411e-02 -2.440112e-02 -2.031367e-02
year2017 8.598713e-02 9.019967e-02 9.426022e-02 9.788880e-02 1.005768e-01
year2018 -8.395387e-03 2.183752e-02 4.779110e-02 6.934389e-02 8.731159e-02
x4046 -1.796573e-07 -1.525806e-07 -1.217385e-07 -8.902625e-08 -5.716638e-08
total_volume -8.333252e-08 -7.587114e-08 -6.872221e-08 -6.185503e-08 -5.535688e-08
large_bags 1.270429e-03 -3.761072e-02 -7.239215e-02 -9.810363e-02 -1.141962e-01
6groups 7groups
type 1.873184e-01 1.810948e-01
regionGreatLakes -6.963509e-02 -5.433530e-02
regionMidsouth -4.189492e-02 -3.019837e-02
regionNortheast 9.728842e-02 1.069021e-01
regionPlains -5.260654e-02 -4.085744e-02
regionSouthCentral -2.231349e-01 -2.233071e-01
regionSoutheast -1.873145e-02 -9.028850e-03
regionWest -4.549297e-02 -3.452304e-02
month2 -3.468319e-02 -3.694705e-02
month3 -2.772264e-04 -2.795779e-03
month4 4.793720e-02 4.876221e-02
month5 5.097630e-02 5.270201e-02
month6 8.569976e-02 8.732103e-02
month7 1.254806e-01 1.282013e-01
month8 1.549898e-01 1.565626e-01
month9 1.900012e-01 1.927043e-01
month10 2.032871e-01 2.064749e-01
month11 1.059677e-01 1.094442e-01
month12 1.643131e-02 2.000957e-02
year2016 -1.673547e-02 -1.371257e-02
year2017 1.022139e-01 1.032881e-01
year2018 1.029213e-01 1.170627e-01
x4046 -2.950483e-08 -9.165685e-09
total_volume -4.949132e-08 -4.476689e-08
large_bags -1.225210e-01 -1.268101e-01AUTOMATED
#### GARBAGE
regsubset_forwards <- regsubsets(average_price ~ .,
data = avocados_prop,
nvmax = 40,
method = "forward")
summary(regsubset_forwards)Subset selection object
Call: regsubsets.formula(average_price ~ ., data = avocados_prop, nvmax = 40,
method = "forward")
29 Variables (and intercept)
Forced in Forced out
total_volume FALSE FALSE
x4046 FALSE FALSE
x4225 FALSE FALSE
x4770 FALSE FALSE
total_bags FALSE FALSE
small_bags FALSE FALSE
large_bags FALSE FALSE
typeorganic FALSE FALSE
year2016 FALSE FALSE
year2017 FALSE FALSE
year2018 FALSE FALSE
regionGreatLakes FALSE FALSE
regionMidsouth FALSE FALSE
regionNortheast FALSE FALSE
regionPlains FALSE FALSE
regionSouthCentral FALSE FALSE
regionSoutheast FALSE FALSE
regionWest FALSE FALSE
month2 FALSE FALSE
month3 FALSE FALSE
month4 FALSE FALSE
month5 FALSE FALSE
month6 FALSE FALSE
month7 FALSE FALSE
month8 FALSE FALSE
month9 FALSE FALSE
month10 FALSE FALSE
month11 FALSE FALSE
month12 FALSE FALSE
1 subsets of each size up to 29
Selection Algorithm: forward
total_volume x4046 x4225 x4770 total_bags small_bags large_bags typeorganic year2016
1 ( 1 ) "*" " " " " " " " " " " " " " " " "
2 ( 1 ) "*" " " " " " " " " " " " " " " " "
3 ( 1 ) "*" " " " " " " " " " " " " " " " "
4 ( 1 ) "*" " " " " " " " " " " " " " " " "
5 ( 1 ) "*" " " " " " " " " " " " " "*" " "
6 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
7 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
8 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
9 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
10 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
11 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
12 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
13 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
14 ( 1 ) "*" " " " " " " " " " " "*" "*" " "
15 ( 1 ) "*" " " "*" " " " " " " "*" "*" " "
16 ( 1 ) "*" " " "*" " " " " " " "*" "*" " "
17 ( 1 ) "*" " " "*" " " " " " " "*" "*" " "
18 ( 1 ) "*" " " "*" " " " " "*" "*" "*" " "
19 ( 1 ) "*" " " "*" " " " " "*" "*" "*" " "
20 ( 1 ) "*" " " "*" " " " " "*" "*" "*" "*"
21 ( 1 ) "*" " " "*" " " " " "*" "*" "*" "*"
22 ( 1 ) "*" " " "*" " " " " "*" "*" "*" "*"
23 ( 1 ) "*" " " "*" " " " " "*" "*" "*" "*"
24 ( 1 ) "*" " " "*" " " " " "*" "*" "*" "*"
25 ( 1 ) "*" "*" "*" " " " " "*" "*" "*" "*"
26 ( 1 ) "*" "*" "*" "*" " " "*" "*" "*" "*"
27 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*" "*"
28 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*" "*"
29 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*" "*"
year2017 year2018 regionGreatLakes regionMidsouth regionNortheast regionPlains
1 ( 1 ) " " " " " " " " " " " "
2 ( 1 ) " " " " " " " " "*" " "
3 ( 1 ) "*" " " " " " " "*" " "
4 ( 1 ) "*" " " " " " " "*" " "
5 ( 1 ) "*" " " " " " " "*" " "
6 ( 1 ) "*" " " " " " " "*" " "
7 ( 1 ) "*" " " " " " " "*" " "
8 ( 1 ) "*" " " " " " " "*" " "
9 ( 1 ) "*" " " " " " " "*" " "
10 ( 1 ) "*" " " " " " " "*" " "
11 ( 1 ) "*" " " " " " " "*" " "
12 ( 1 ) "*" "*" " " " " "*" " "
13 ( 1 ) "*" "*" " " " " "*" " "
14 ( 1 ) "*" "*" " " " " "*" " "
15 ( 1 ) "*" "*" " " " " "*" " "
16 ( 1 ) "*" "*" " " " " "*" " "
17 ( 1 ) "*" "*" " " " " "*" " "
18 ( 1 ) "*" "*" " " " " "*" " "
19 ( 1 ) "*" "*" "*" " " "*" " "
20 ( 1 ) "*" "*" "*" " " "*" " "
21 ( 1 ) "*" "*" "*" " " "*" " "
22 ( 1 ) "*" "*" "*" " " "*" " "
23 ( 1 ) "*" "*" "*" " " "*" "*"
24 ( 1 ) "*" "*" "*" "*" "*" "*"
25 ( 1 ) "*" "*" "*" "*" "*" "*"
26 ( 1 ) "*" "*" "*" "*" "*" "*"
27 ( 1 ) "*" "*" "*" "*" "*" "*"
28 ( 1 ) "*" "*" "*" "*" "*" "*"
29 ( 1 ) "*" "*" "*" "*" "*" "*"
regionSouthCentral regionSoutheast regionWest month2 month3 month4 month5 month6 month7
1 ( 1 ) " " " " " " " " " " " " " " " " " "
2 ( 1 ) " " " " " " " " " " " " " " " " " "
3 ( 1 ) " " " " " " " " " " " " " " " " " "
4 ( 1 ) "*" " " " " " " " " " " " " " " " "
5 ( 1 ) "*" " " " " " " " " " " " " " " " "
6 ( 1 ) "*" " " " " " " " " " " " " " " " "
7 ( 1 ) "*" " " " " " " " " " " " " " " " "
8 ( 1 ) "*" " " " " " " " " " " " " " " " "
9 ( 1 ) "*" " " " " " " " " " " " " " " " "
10 ( 1 ) "*" " " " " " " " " " " " " " " "*"
11 ( 1 ) "*" " " " " " " " " " " " " " " "*"
12 ( 1 ) "*" " " " " " " " " " " " " " " "*"
13 ( 1 ) "*" " " " " " " " " " " " " "*" "*"
14 ( 1 ) "*" " " " " "*" " " " " " " "*" "*"
15 ( 1 ) "*" " " " " "*" " " " " " " "*" "*"
16 ( 1 ) "*" " " " " "*" " " " " "*" "*" "*"
17 ( 1 ) "*" " " " " "*" " " "*" "*" "*" "*"
18 ( 1 ) "*" " " " " "*" " " "*" "*" "*" "*"
19 ( 1 ) "*" " " " " "*" " " "*" "*" "*" "*"
20 ( 1 ) "*" " " " " "*" " " "*" "*" "*" "*"
21 ( 1 ) "*" " " " " "*" " " "*" "*" "*" "*"
22 ( 1 ) "*" " " "*" "*" " " "*" "*" "*" "*"
23 ( 1 ) "*" " " "*" "*" " " "*" "*" "*" "*"
24 ( 1 ) "*" " " "*" "*" " " "*" "*" "*" "*"
25 ( 1 ) "*" " " "*" "*" " " "*" "*" "*" "*"
26 ( 1 ) "*" " " "*" "*" " " "*" "*" "*" "*"
27 ( 1 ) "*" " " "*" "*" " " "*" "*" "*" "*"
28 ( 1 ) "*" "*" "*" "*" " " "*" "*" "*" "*"
29 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*" "*"
month8 month9 month10 month11 month12
1 ( 1 ) " " " " " " " " " "
2 ( 1 ) " " " " " " " " " "
3 ( 1 ) " " " " " " " " " "
4 ( 1 ) " " " " " " " " " "
5 ( 1 ) " " " " " " " " " "
6 ( 1 ) " " " " " " " " " "
7 ( 1 ) " " " " "*" " " " "
8 ( 1 ) " " "*" "*" " " " "
9 ( 1 ) "*" "*" "*" " " " "
10 ( 1 ) "*" "*" "*" " " " "
11 ( 1 ) "*" "*" "*" "*" " "
12 ( 1 ) "*" "*" "*" "*" " "
13 ( 1 ) "*" "*" "*" "*" " "
14 ( 1 ) "*" "*" "*" "*" " "
15 ( 1 ) "*" "*" "*" "*" " "
16 ( 1 ) "*" "*" "*" "*" " "
17 ( 1 ) "*" "*" "*" "*" " "
18 ( 1 ) "*" "*" "*" "*" " "
19 ( 1 ) "*" "*" "*" "*" " "
20 ( 1 ) "*" "*" "*" "*" " "
21 ( 1 ) "*" "*" "*" "*" "*"
22 ( 1 ) "*" "*" "*" "*" "*"
23 ( 1 ) "*" "*" "*" "*" "*"
24 ( 1 ) "*" "*" "*" "*" "*"
25 ( 1 ) "*" "*" "*" "*" "*"
26 ( 1 ) "*" "*" "*" "*" "*"
27 ( 1 ) "*" "*" "*" "*" "*"
28 ( 1 ) "*" "*" "*" "*" "*"
29 ( 1 ) "*" "*" "*" "*" "*" plot(regsubset_forwards,
scale = "bic")
sum_forward <- summary(regsubset_forwards)
plot(sum_forward$bic,
type = "b")
# glmulti
glmulti_fit <- glmulti(
log(average_price) ~ ., # model to fit, in this case, charges varies with everything
level = 2, # level = 2 means try pairwise interactions. level = 1 means main effects only
data = train, # data to use for fitting
minsize = 0, # min size of model to try, in number of predictors
maxsize = -1, # max size to try, set to -1 for unlimited
marginality = TRUE, # marginality true means include pairwise interaction only if both main effects present in model.
method = "d", # method "d" means trial run, to get size of problem. Set to "h" for exhaustive search, or "g" for genetic algorithm
confsetsize = 10, # how many models should glmulti() return? Must be less than total size of problem
plotty = FALSE, # provide progress plots? Generally annoying.
report = TRUE, # provide progress reports? Generally useful.
fitfunction = lm, # use lm() as fit function. Can also use glm() for logistic regression.
crit = aic # criterion for selecting best models.
)Initialization...
TASK: Diagnostic of candidate set.
Sample size: 2164
4 factor(s).
7 covariate(s).
0 f exclusion(s).
0 c exclusion(s).
0 f:f exclusion(s).
0 c:c exclusion(s).
0 f:c exclusion(s).
Size constraints: min = 0 max = -1
Complexity constraints: min = 0 max = -1
Marginality rule.
Your candidate set contains more than 1 billion (1e9) models.lm_multi <- lm(log(average_price)~1+type+year+region+month+total_volume+x4046+x4225+x4770+total_bags+small_bags+large_bags+year:type+region:type+region:year+month:type+month:year+month:region+x4046:total_volume+x4225:total_volume+total_bags:x4225+total_bags:x4770+small_bags:total_volume+small_bags:x4046+small_bags:x4225+small_bags:x4770+small_bags:total_bags+large_bags:total_volume+large_bags:small_bags+type:total_volume+type:x4225+type:total_bags+year:total_volume+year:x4046+year:x4770+year:total_bags+year:large_bags+region:total_volume+region:x4225+region:total_bags+region:large_bags+month:total_volume+month:x4770+month:small_bags, train)
summary(lm_multi)
Call:
lm(formula = log(average_price) ~ 1 + type + year + region +
month + total_volume + x4046 + x4225 + x4770 + total_bags +
small_bags + large_bags + year:type + region:type + region:year +
month:type + month:year + month:region + x4046:total_volume +
x4225:total_volume + total_bags:x4225 + total_bags:x4770 +
small_bags:total_volume + small_bags:x4046 + small_bags:x4225 +
small_bags:x4770 + small_bags:total_bags + large_bags:total_volume +
large_bags:small_bags + type:total_volume + type:x4225 +
type:total_bags + year:total_volume + year:x4046 + year:x4770 +
year:total_bags + year:large_bags + region:total_volume +
region:x4225 + region:total_bags + region:large_bags + month:total_volume +
month:x4770 + month:small_bags, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.221001 -0.032174 0.001345 0.033539 0.192423
Coefficients: (9 not defined because of singularities)
Estimate Std. Error t value Pr(>|t|)
(Intercept) 8.165e-03 4.094e-01 0.020 0.984090
typeorganic -2.784e-01 7.165e-02 -3.886 0.000105 ***
year2016 2.264e-01 2.893e-02 7.825 8.33e-15 ***
year2017 2.530e-01 2.917e-02 8.673 < 2e-16 ***
year2018 5.585e-01 4.375e-02 12.765 < 2e-16 ***
regionGreatLakes -1.878e-01 6.510e-02 -2.884 0.003970 **
regionMidsouth -3.676e-01 6.743e-02 -5.451 5.67e-08 ***
regionNortheast -1.116e-01 6.981e-02 -1.598 0.110148
regionPlains -3.348e-01 7.045e-02 -4.752 2.17e-06 ***
regionSouthCentral -2.783e-01 6.355e-02 -4.379 1.26e-05 ***
regionSoutheast -2.106e-01 6.151e-02 -3.423 0.000632 ***
regionWest -1.405e-01 5.894e-02 -2.383 0.017256 *
month2 4.356e-02 5.527e-02 0.788 0.430685
month3 9.905e-02 5.560e-02 1.781 0.075010 .
month4 1.934e-01 5.974e-02 3.237 0.001230 **
month5 1.794e-01 6.119e-02 2.931 0.003416 **
month6 2.517e-01 7.364e-02 3.419 0.000643 ***
month7 2.664e-01 6.506e-02 4.094 4.41e-05 ***
month8 3.411e-01 6.150e-02 5.546 3.33e-08 ***
month9 1.978e-01 6.144e-02 3.219 0.001307 **
month10 2.245e-01 5.838e-02 3.845 0.000124 ***
month11 8.393e-02 5.889e-02 1.425 0.154253
month12 -1.671e-02 5.856e-02 -0.285 0.775332
total_volume 3.166e-04 1.371e-04 2.309 0.021043 *
x4046 -3.165e-04 1.371e-04 -2.308 0.021079 *
x4225 -3.167e-04 1.371e-04 -2.310 0.021003 *
x4770 -3.166e-04 1.371e-04 -2.310 0.021017 *
total_bags -3.166e-04 1.371e-04 -2.309 0.021060 *
small_bags 7.795e-01 4.104e-01 1.899 0.057662 .
large_bags 5.540e-01 4.177e-01 1.326 0.184886
typeorganic:year2016 -1.797e-01 2.226e-02 -8.070 1.24e-15 ***
typeorganic:year2017 -1.683e-01 2.358e-02 -7.138 1.34e-12 ***
typeorganic:year2018 -1.956e-01 3.689e-02 -5.303 1.27e-07 ***
typeorganic:regionGreatLakes 2.345e-01 6.041e-02 3.883 0.000107 ***
typeorganic:regionMidsouth 4.407e-01 6.377e-02 6.911 6.54e-12 ***
typeorganic:regionNortheast 3.100e-01 6.615e-02 4.687 2.97e-06 ***
typeorganic:regionPlains 4.703e-01 6.813e-02 6.903 6.92e-12 ***
typeorganic:regionSouthCentral 1.715e-01 6.020e-02 2.848 0.004442 **
typeorganic:regionSoutheast 3.349e-01 5.767e-02 5.807 7.43e-09 ***
typeorganic:regionWest 3.563e-01 4.709e-02 7.566 5.96e-14 ***
year2016:regionGreatLakes -9.946e-02 2.185e-02 -4.551 5.67e-06 ***
year2017:regionGreatLakes -5.205e-02 2.139e-02 -2.433 0.015047 *
year2018:regionGreatLakes -1.579e-01 3.763e-02 -4.196 2.85e-05 ***
year2016:regionMidsouth -4.469e-02 1.759e-02 -2.541 0.011140 *
year2017:regionMidsouth -2.629e-03 1.684e-02 -0.156 0.875938
year2018:regionMidsouth -1.130e-01 2.918e-02 -3.874 0.000111 ***
year2016:regionNortheast 4.364e-02 1.699e-02 2.568 0.010292 *
year2017:regionNortheast 9.924e-02 1.700e-02 5.837 6.24e-09 ***
year2018:regionNortheast 6.734e-02 3.189e-02 2.112 0.034848 *
year2016:regionPlains -1.492e-01 1.919e-02 -7.775 1.22e-14 ***
year2017:regionPlains -1.797e-01 1.877e-02 -9.572 < 2e-16 ***
year2018:regionPlains -3.886e-01 3.472e-02 -11.193 < 2e-16 ***
year2016:regionSouthCentral -8.497e-02 1.643e-02 -5.172 2.57e-07 ***
year2017:regionSouthCentral -7.938e-02 1.558e-02 -5.094 3.85e-07 ***
year2018:regionSouthCentral -1.206e-01 2.916e-02 -4.136 3.68e-05 ***
year2016:regionSoutheast -1.606e-01 1.823e-02 -8.807 < 2e-16 ***
year2017:regionSoutheast -1.263e-01 1.784e-02 -7.079 2.04e-12 ***
year2018:regionSoutheast -3.494e-01 3.109e-02 -11.237 < 2e-16 ***
year2016:regionWest -1.693e-01 2.585e-02 -6.551 7.36e-11 ***
year2017:regionWest -7.295e-02 2.652e-02 -2.751 0.005993 **
year2018:regionWest -2.010e-01 4.993e-02 -4.026 5.89e-05 ***
typeorganic:month2 -3.940e-02 2.835e-02 -1.390 0.164707
typeorganic:month3 -7.479e-02 2.903e-02 -2.576 0.010058 *
typeorganic:month4 -4.909e-02 3.198e-02 -1.535 0.124877
typeorganic:month5 -6.119e-02 3.216e-02 -1.903 0.057238 .
typeorganic:month6 -5.747e-02 3.696e-02 -1.555 0.120154
typeorganic:month7 -4.697e-02 3.410e-02 -1.377 0.168534
typeorganic:month8 -4.344e-02 3.278e-02 -1.325 0.185280
typeorganic:month9 4.129e-02 3.183e-02 1.297 0.194727
typeorganic:month10 -7.466e-02 3.085e-02 -2.420 0.015603 *
typeorganic:month11 1.555e-02 3.114e-02 0.499 0.617628
typeorganic:month12 6.876e-02 3.062e-02 2.246 0.024828 *
year2016:month2 1.078e-02 1.656e-02 0.651 0.515194
year2017:month2 1.737e-02 1.589e-02 1.093 0.274372
year2018:month2 -5.682e-03 1.637e-02 -0.347 0.728638
year2016:month3 1.382e-02 1.661e-02 0.832 0.405684
year2017:month3 1.186e-01 1.604e-02 7.393 2.14e-13 ***
year2018:month3 2.810e-03 1.621e-02 0.173 0.862403
year2016:month4 -5.759e-02 1.623e-02 -3.547 0.000399 ***
year2017:month4 1.362e-01 1.554e-02 8.765 < 2e-16 ***
year2018:month4 NA NA NA NA
year2016:month5 -4.515e-02 1.548e-02 -2.917 0.003579 **
year2017:month5 1.960e-01 1.566e-02 12.520 < 2e-16 ***
year2018:month5 NA NA NA NA
year2016:month6 -1.890e-02 1.669e-02 -1.132 0.257598
year2017:month6 1.659e-01 1.672e-02 9.922 < 2e-16 ***
year2018:month6 NA NA NA NA
year2016:month7 5.033e-02 1.611e-02 3.125 0.001805 **
year2017:month7 1.556e-01 1.568e-02 9.919 < 2e-16 ***
year2018:month7 NA NA NA NA
year2016:month8 4.055e-02 1.607e-02 2.524 0.011692 *
year2017:month8 2.032e-01 1.576e-02 12.894 < 2e-16 ***
year2018:month8 NA NA NA NA
year2016:month9 5.450e-02 1.695e-02 3.215 0.001326 **
year2017:month9 2.789e-01 1.592e-02 17.516 < 2e-16 ***
year2018:month9 NA NA NA NA
year2016:month10 1.130e-01 1.710e-02 6.605 5.13e-11 ***
year2017:month10 2.422e-01 1.659e-02 14.598 < 2e-16 ***
year2018:month10 NA NA NA NA
year2016:month11 1.736e-01 1.648e-02 10.534 < 2e-16 ***
year2017:month11 1.820e-01 1.598e-02 11.395 < 2e-16 ***
year2018:month11 NA NA NA NA
year2016:month12 3.590e-02 1.671e-02 2.149 0.031772 *
year2017:month12 1.342e-01 1.647e-02 8.145 6.79e-16 ***
year2018:month12 NA NA NA NA
regionGreatLakes:month2 -5.116e-02 2.694e-02 -1.899 0.057724 .
regionMidsouth:month2 -5.640e-02 2.445e-02 -2.306 0.021201 *
regionNortheast:month2 1.266e-02 2.342e-02 0.540 0.588972
regionPlains:month2 -6.890e-02 2.673e-02 -2.577 0.010039 *
regionSouthCentral:month2 -1.542e-02 2.203e-02 -0.700 0.484020
regionSoutheast:month2 -5.687e-02 2.592e-02 -2.194 0.028341 *
regionWest:month2 -1.872e-03 2.706e-02 -0.069 0.944864
regionGreatLakes:month3 -5.322e-02 2.686e-02 -1.981 0.047714 *
regionMidsouth:month3 -7.974e-02 2.378e-02 -3.353 0.000815 ***
regionNortheast:month3 -4.667e-02 2.269e-02 -2.057 0.039843 *
regionPlains:month3 -8.713e-02 2.599e-02 -3.352 0.000818 ***
regionSouthCentral:month3 -3.027e-02 2.208e-02 -1.371 0.170585
regionSoutheast:month3 -8.725e-02 2.493e-02 -3.499 0.000477 ***
regionWest:month3 -6.355e-03 2.644e-02 -0.240 0.810116
regionGreatLakes:month4 -7.740e-02 2.934e-02 -2.638 0.008396 **
regionMidsouth:month4 -7.437e-02 2.648e-02 -2.808 0.005032 **
regionNortheast:month4 -2.576e-02 2.511e-02 -1.026 0.305047
regionPlains:month4 -9.680e-02 2.879e-02 -3.362 0.000788 ***
regionSouthCentral:month4 -3.464e-02 2.391e-02 -1.449 0.147579
regionSoutheast:month4 -8.844e-02 2.684e-02 -3.295 0.001002 **
regionWest:month4 -5.382e-02 2.866e-02 -1.878 0.060526 .
regionGreatLakes:month5 -5.397e-02 3.033e-02 -1.779 0.075371 .
regionMidsouth:month5 -7.007e-02 2.705e-02 -2.591 0.009649 **
regionNortheast:month5 -1.297e-03 2.491e-02 -0.052 0.958465
regionPlains:month5 -7.402e-02 2.900e-02 -2.552 0.010787 *
regionSouthCentral:month5 -1.575e-02 2.442e-02 -0.645 0.518930
regionSoutheast:month5 -7.682e-02 2.765e-02 -2.779 0.005512 **
regionWest:month5 3.702e-03 2.895e-02 0.128 0.898251
regionGreatLakes:month6 -1.351e-01 3.620e-02 -3.733 0.000195 ***
regionMidsouth:month6 -1.108e-01 3.067e-02 -3.613 0.000310 ***
regionNortheast:month6 -4.064e-02 2.689e-02 -1.511 0.130916
regionPlains:month6 -1.275e-01 3.184e-02 -4.003 6.48e-05 ***
regionSouthCentral:month6 -7.873e-02 2.599e-02 -3.030 0.002482 **
regionSoutheast:month6 -1.451e-01 3.113e-02 -4.662 3.35e-06 ***
regionWest:month6 -4.059e-02 3.423e-02 -1.186 0.235943
regionGreatLakes:month7 -1.250e-01 3.315e-02 -3.772 0.000167 ***
regionMidsouth:month7 -1.357e-01 2.801e-02 -4.845 1.37e-06 ***
regionNortheast:month7 -1.183e-01 2.582e-02 -4.583 4.87e-06 ***
regionPlains:month7 -1.613e-01 2.958e-02 -5.453 5.60e-08 ***
regionSouthCentral:month7 -9.911e-02 2.399e-02 -4.131 3.77e-05 ***
regionSoutheast:month7 -1.978e-01 2.753e-02 -7.182 9.79e-13 ***
regionWest:month7 -3.284e-02 3.053e-02 -1.076 0.282137
regionGreatLakes:month8 -1.541e-01 3.085e-02 -4.994 6.44e-07 ***
regionMidsouth:month8 -1.473e-01 2.679e-02 -5.500 4.30e-08 ***
regionNortheast:month8 -2.010e-01 2.694e-02 -7.460 1.30e-13 ***
regionPlains:month8 -1.741e-01 2.998e-02 -5.809 7.36e-09 ***
regionSouthCentral:month8 -8.039e-02 2.503e-02 -3.211 0.001344 **
regionSoutheast:month8 -2.004e-01 2.802e-02 -7.154 1.20e-12 ***
regionWest:month8 -7.251e-02 2.955e-02 -2.454 0.014216 *
regionGreatLakes:month9 -8.688e-02 3.022e-02 -2.875 0.004083 **
regionMidsouth:month9 -9.526e-02 2.731e-02 -3.488 0.000498 ***
regionNortheast:month9 -1.451e-01 2.701e-02 -5.374 8.64e-08 ***
regionPlains:month9 -1.061e-01 2.987e-02 -3.550 0.000394 ***
regionSouthCentral:month9 -3.826e-02 2.492e-02 -1.536 0.124780
regionSoutheast:month9 -1.027e-01 2.789e-02 -3.681 0.000239 ***
regionWest:month9 -6.269e-02 3.036e-02 -2.065 0.039054 *
regionGreatLakes:month10 -1.015e-01 2.997e-02 -3.388 0.000718 ***
regionMidsouth:month10 -1.333e-01 2.665e-02 -5.003 6.16e-07 ***
regionNortheast:month10 -1.992e-01 2.537e-02 -7.854 6.71e-15 ***
regionPlains:month10 -1.700e-01 2.888e-02 -5.884 4.71e-09 ***
regionSouthCentral:month10 -3.319e-02 2.386e-02 -1.391 0.164348
regionSoutheast:month10 -1.208e-01 2.772e-02 -4.359 1.38e-05 ***
regionWest:month10 -4.649e-02 2.989e-02 -1.555 0.120058
regionGreatLakes:month11 -1.109e-01 3.057e-02 -3.627 0.000294 ***
regionMidsouth:month11 -1.030e-01 2.679e-02 -3.846 0.000124 ***
regionNortheast:month11 -1.235e-01 2.561e-02 -4.822 1.53e-06 ***
regionPlains:month11 -9.488e-02 2.963e-02 -3.203 0.001385 **
regionSouthCentral:month11 -1.026e-02 2.487e-02 -0.412 0.680117
regionSoutheast:month11 -1.334e-01 2.761e-02 -4.832 1.46e-06 ***
regionWest:month11 -4.012e-02 3.018e-02 -1.330 0.183808
regionGreatLakes:month12 -4.482e-02 2.890e-02 -1.551 0.121157
regionMidsouth:month12 -4.076e-02 2.610e-02 -1.562 0.118516
regionNortheast:month12 1.524e-04 2.512e-02 0.006 0.995159
regionPlains:month12 -5.109e-02 2.847e-02 -1.795 0.072890 .
regionSouthCentral:month12 1.416e-02 2.346e-02 0.603 0.546279
regionSoutheast:month12 -9.327e-02 2.770e-02 -3.367 0.000775 ***
regionWest:month12 -1.491e-02 3.144e-02 -0.474 0.635416
total_volume:x4046 9.095e-15 3.453e-15 2.634 0.008516 **
total_volume:x4225 1.971e-14 4.402e-15 4.476 8.05e-06 ***
x4225:total_bags 3.843e-14 1.085e-14 3.542 0.000406 ***
x4770:total_bags -1.550e-13 6.747e-14 -2.298 0.021694 *
total_volume:small_bags -4.496e-04 2.089e-04 -2.153 0.031455 *
x4046:small_bags 4.494e-04 2.089e-04 2.152 0.031558 *
x4225:small_bags 4.495e-04 2.089e-04 2.152 0.031518 *
x4770:small_bags 4.497e-04 2.088e-04 2.153 0.031432 *
total_bags:small_bags 4.493e-04 2.089e-04 2.151 0.031570 *
total_volume:large_bags -2.281e-07 1.031e-07 -2.213 0.027020 *
small_bags:large_bags 2.682e-01 5.652e-02 4.744 2.25e-06 ***
typeorganic:total_volume -1.877e-06 1.644e-07 -11.414 < 2e-16 ***
typeorganic:x4225 2.049e-06 2.275e-07 9.007 < 2e-16 ***
typeorganic:total_bags -5.807e-07 1.743e-07 -3.331 0.000883 ***
year2016:total_volume -1.197e-07 1.125e-08 -10.640 < 2e-16 ***
year2017:total_volume -1.098e-07 1.171e-08 -9.375 < 2e-16 ***
year2018:total_volume -1.407e-07 1.440e-08 -9.770 < 2e-16 ***
year2016:x4046 8.473e-08 1.103e-08 7.683 2.47e-14 ***
[ reached getOption("max.print") -- omitted 72 rows ]
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.05349 on 1901 degrees of freedom
Multiple R-squared: 0.9642, Adjusted R-squared: 0.9593
F-statistic: 195.4 on 262 and 1901 DF, p-value: < 2.2e-16glance(lm_multi)train_rmse <- train %>%
add_residuals(lm_multi) %>%
mutate(sq_resid = resid^2) %>%
summarise(mse = mean(sq_resid),
rmse = mse^0.5) %>%
pull(rmse)Warning: prediction from a rank-deficient fit may be misleadingtrain_rmse[1] 0.05013213predictions_test <- test %>%
add_predictions(lm_multi) %>%
add_residuals(lm_multi) %>%
dplyr::select(average_price, pred, resid) Warning: prediction from a rank-deficient fit may be misleadingWarning: prediction from a rank-deficient fit may be misleading
test_rsme <- predictions_test %>%
mutate(sq_resid = resid^2) %>%
summarise(mse = mean(sq_resid),
rmse = mse^0.5) %>%
pull(rmse)
test_rsme / train_rmse[1] 1.265112# over fit? despite having a good bicglmulti_fit <- glmulti(
log(average_price) ~ ., # model to fit, in this case, charges varies with everything
level = 2, # level = 2 means try pairwise interactions. level = 1 means main effects only
data = train, # data to use for fitting
minsize = 0, # min size of model to try, in number of predictors
maxsize = 10, # max size to try, set to -1 for unlimited
marginality = TRUE, # marginality true means include pairwise interaction only if both main effects present in model.
method = "d", # method "d" means trial run, to get size of problem. Set to "h" for exhaustive search, or "g" for genetic algorithm
confsetsize = 1000, # how many models should glmulti() return? Must be less than total size of problem
plotty = FALSE, # provide progress plots? Generally annoying.
report = TRUE, # provide progress reports? Generally useful.
fitfunction = lm, # use lm() as fit function. Can also use glm() for logistic regression.
crit = aic # criterion for selecting best models.
)train = train %>%
select(average_price, region, type, year, month, total_volume, large_bags)
lm_multi <- lm(log(average_price)~ 1 + region + type + year + month + total_volume + large_bags + year:type + month:year + region:total_volume + type:total_volume, train)
summary(lm_multi)
Call:
lm(formula = log(average_price) ~ 1 + region + type + year +
month + total_volume + large_bags + year:type + month:year +
region:total_volume + type:total_volume, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.34697 -0.04625 0.00192 0.04766 0.53098
Coefficients: (9 not defined because of singularities)
Estimate Std. Error t value Pr(>|t|)
(Intercept) 4.663e-01 1.828e-02 25.506 < 2e-16 ***
regionGreatLakes -5.460e-02 1.121e-02 -4.871 1.19e-06 ***
regionMidsouth -5.180e-02 1.038e-02 -4.993 6.45e-07 ***
regionNortheast 1.023e-01 9.396e-03 10.884 < 2e-16 ***
regionPlains -4.999e-02 1.141e-02 -4.381 1.24e-05 ***
regionSouthCentral -2.804e-01 9.669e-03 -28.998 < 2e-16 ***
regionSoutheast -1.015e-01 1.085e-02 -9.358 < 2e-16 ***
regionWest 1.086e-01 1.271e-02 8.546 < 2e-16 ***
typeorganic 1.814e-01 1.465e-02 12.383 < 2e-16 ***
year2016 4.252e-02 1.509e-02 2.819 0.004869 **
year2017 1.042e-01 1.552e-02 6.716 2.40e-11 ***
year2018 2.433e-01 1.705e-02 14.265 < 2e-16 ***
month2 8.632e-03 1.556e-02 0.555 0.579028
month3 1.649e-02 1.472e-02 1.120 0.262703
month4 6.147e-02 1.530e-02 4.017 6.10e-05 ***
month5 5.158e-02 1.497e-02 3.446 0.000580 ***
month6 7.818e-02 1.534e-02 5.095 3.79e-07 ***
month7 8.055e-02 1.565e-02 5.148 2.88e-07 ***
month8 8.920e-02 1.428e-02 6.246 5.07e-10 ***
month9 6.977e-02 1.536e-02 4.543 5.86e-06 ***
month10 3.473e-02 1.554e-02 2.235 0.025536 *
month11 -2.938e-02 1.468e-02 -2.001 0.045483 *
month12 -3.794e-02 1.494e-02 -2.538 0.011206 *
total_volume -8.671e-08 2.850e-09 -30.428 < 2e-16 ***
large_bags -1.888e-01 1.366e-02 -13.821 < 2e-16 ***
typeorganic:year2016 -8.678e-02 8.663e-03 -10.018 < 2e-16 ***
typeorganic:year2017 -1.263e-01 9.171e-03 -13.767 < 2e-16 ***
typeorganic:year2018 -1.221e-01 1.564e-02 -7.810 8.92e-15 ***
year2016:month2 -1.740e-02 2.082e-02 -0.836 0.403495
year2017:month2 -3.230e-02 2.114e-02 -1.528 0.126740
year2018:month2 -2.678e-02 2.166e-02 -1.236 0.216488
year2016:month3 -1.764e-02 2.023e-02 -0.872 0.383462
year2017:month3 5.821e-02 2.060e-02 2.825 0.004766 **
year2018:month3 -3.319e-02 2.096e-02 -1.583 0.113533
year2016:month4 -6.828e-02 2.077e-02 -3.288 0.001025 **
year2017:month4 1.033e-01 2.047e-02 5.046 4.89e-07 ***
year2018:month4 NA NA NA NA
year2016:month5 -3.712e-02 2.012e-02 -1.845 0.065210 .
year2017:month5 1.571e-01 2.091e-02 7.512 8.57e-14 ***
year2018:month5 NA NA NA NA
year2016:month6 -2.374e-02 2.067e-02 -1.149 0.250864
year2017:month6 1.148e-01 2.102e-02 5.463 5.23e-08 ***
year2018:month6 NA NA NA NA
year2016:month7 4.701e-02 2.073e-02 2.268 0.023445 *
year2017:month7 1.328e-01 2.112e-02 6.286 3.95e-10 ***
year2018:month7 NA NA NA NA
year2016:month8 2.643e-02 2.014e-02 1.312 0.189538
year2017:month8 1.663e-01 2.044e-02 8.135 6.95e-16 ***
year2018:month8 NA NA NA NA
year2016:month9 6.379e-02 2.077e-02 3.071 0.002159 **
year2017:month9 2.513e-01 2.147e-02 11.706 < 2e-16 ***
year2018:month9 NA NA NA NA
year2016:month10 1.389e-01 2.053e-02 6.766 1.72e-11 ***
year2017:month10 2.413e-01 2.104e-02 11.464 < 2e-16 ***
year2018:month10 NA NA NA NA
year2016:month11 1.972e-01 2.048e-02 9.626 < 2e-16 ***
year2017:month11 1.663e-01 2.074e-02 8.017 1.77e-15 ***
year2018:month11 NA NA NA NA
year2016:month12 3.177e-02 2.053e-02 1.547 0.121906
year2017:month12 1.024e-01 2.050e-02 4.998 6.29e-07 ***
year2018:month12 NA NA NA NA
regionGreatLakes:total_volume -2.009e-08 3.755e-09 -5.350 9.78e-08 ***
regionMidsouth:total_volume -3.521e-08 4.501e-09 -7.822 8.14e-15 ***
regionNortheast:total_volume -9.684e-09 2.902e-09 -3.337 0.000861 ***
regionPlains:total_volume -1.384e-07 7.841e-09 -17.646 < 2e-16 ***
regionSouthCentral:total_volume 1.061e-08 2.251e-09 4.711 2.62e-06 ***
regionSoutheast:total_volume 1.130e-09 3.422e-09 0.330 0.741157
regionWest:total_volume -2.190e-08 2.301e-09 -9.519 < 2e-16 ***
typeorganic:total_volume -1.213e-06 4.779e-08 -25.372 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.07631 on 2104 degrees of freedom
Multiple R-squared: 0.9171, Adjusted R-squared: 0.9148
F-statistic: 394.6 on 59 and 2104 DF, p-value: < 2.2e-16glance(lm_multi)plot(lm_multi)
lm12 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume + type:region + type:year + region:total_volume, train)
summary(lm12)
Call:
lm(formula = log(average_price) ~ type + region + month + year +
x4046 + total_volume + large_bags + type:total_volume + type:region +
type:year + region:total_volume, data = train)
Residuals:
Min 1Q Median 3Q Max
-0.29711 -0.05424 -0.00103 0.05371 0.58560
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 5.040e-01 4.200e-02 12.000 < 2e-16 ***
typeorganic 1.132e-01 4.174e-02 2.712 0.006737 **
regionGreatLakes 4.623e-02 5.474e-02 0.845 0.398458
regionMidsouth -6.778e-02 5.790e-02 -1.171 0.241853
regionNortheast -9.002e-02 5.419e-02 -1.661 0.096845 .
regionPlains -1.949e-02 5.207e-02 -0.374 0.708235
regionSouthCentral -1.663e-01 6.121e-02 -2.717 0.006641 **
regionSoutheast -6.441e-02 5.020e-02 -1.283 0.199627
regionWest -1.099e-01 5.461e-02 -2.013 0.044280 *
month2 -5.992e-03 8.212e-03 -0.730 0.465728
month3 2.174e-02 8.023e-03 2.710 0.006777 **
month4 7.506e-02 8.803e-03 8.527 < 2e-16 ***
month5 9.300e-02 8.875e-03 10.479 < 2e-16 ***
month6 1.089e-01 8.942e-03 12.181 < 2e-16 ***
month7 1.411e-01 8.794e-03 16.051 < 2e-16 ***
month8 1.452e-01 8.775e-03 16.547 < 2e-16 ***
month9 1.671e-01 9.052e-03 18.459 < 2e-16 ***
month10 1.585e-01 8.883e-03 17.847 < 2e-16 ***
month11 7.760e-02 9.009e-03 8.614 < 2e-16 ***
month12 2.495e-03 8.786e-03 0.284 0.776443
year2016 9.240e-02 7.159e-03 12.907 < 2e-16 ***
year2017 2.423e-01 6.958e-03 34.822 < 2e-16 ***
year2018 3.110e-01 1.221e-02 25.475 < 2e-16 ***
x4046 4.992e-08 9.840e-09 5.073 4.26e-07 ***
total_volume -1.229e-07 7.607e-09 -16.158 < 2e-16 ***
large_bags -2.163e-01 1.476e-02 -14.657 < 2e-16 ***
typeorganic:total_volume -1.338e-06 5.432e-08 -24.634 < 2e-16 ***
typeorganic:regionGreatLakes -9.084e-02 5.398e-02 -1.683 0.092572 .
typeorganic:regionMidsouth 1.609e-02 5.713e-02 0.282 0.778277
typeorganic:regionNortheast 1.904e-01 5.361e-02 3.551 0.000393 ***
typeorganic:regionPlains -3.563e-02 5.227e-02 -0.682 0.495506
typeorganic:regionSouthCentral -1.245e-01 6.087e-02 -2.046 0.040916 *
typeorganic:regionSoutheast -4.898e-02 5.038e-02 -0.972 0.330996
typeorganic:regionWest 2.467e-01 5.423e-02 4.549 5.70e-06 ***
typeorganic:year2016 -9.931e-02 9.931e-03 -10.000 < 2e-16 ***
typeorganic:year2017 -1.291e-01 1.028e-02 -12.552 < 2e-16 ***
typeorganic:year2018 -1.448e-01 1.733e-02 -8.356 < 2e-16 ***
regionGreatLakes:total_volume -4.966e-08 1.280e-08 -3.880 0.000108 ***
regionMidsouth:total_volume -3.889e-08 1.574e-08 -2.471 0.013540 *
regionNortheast:total_volume 4.550e-08 1.119e-08 4.066 4.95e-05 ***
regionPlains:total_volume -1.954e-07 1.911e-08 -10.229 < 2e-16 ***
regionSouthCentral:total_volume -1.441e-08 1.029e-08 -1.401 0.161390
regionSoutheast:total_volume -2.368e-08 1.060e-08 -2.233 0.025684 *
regionWest:total_volume 1.558e-08 8.890e-09 1.753 0.079786 .
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.08421 on 2120 degrees of freedom
Multiple R-squared: 0.8983, Adjusted R-squared: 0.8962
F-statistic: 435.4 on 43 and 2120 DF, p-value: < 2.2e-16glance(lm12)plot(lm12)
---
title: "R Notebook"
output: html_notebook
---

```{r}
library(tidyverse)
library(lubridate)
library(janitor)
library(caret)
library(broom)
library(fastDummies)
library(GGally)
library(ggfortify)
library(mosaic)
library(mosaicData)
library(modelr)
library(relaimpo)
library(tidyverse)
library(glmulti)
```

```{r}
avocados <- read_csv(here::here("weekend/data/avocado.csv")) %>% clean_names()
```

```{r}
avocados %>%
  distinct(region)

avocados %>% 
  filter(region == "TotalUS",
         date == "2015-12-27")

avocados %>% 
  filter(region %in% c("Midsouth", "Northeast", "Plains", "SouthCentral", "Southeast", "West", "GreatLakes", "California"),
         date == "2015-12-27") %>% 
  group_by(type) %>% 
  summarise(total_volume = sum(total_volume),
            x4046 = sum(x4046),
            total_bags = sum(total_bags))

avocados %>% 
  filter(!region %in% c("TotalUS", "Midsouth", "Northeast", "Plains", "SouthCentral", "Southeast", "West", "GreatLakes", "California"),
         date == "2015-12-27") %>% 
  group_by(type) %>% 
  summarise(total_volume = sum(total_volume),
            x4046 = sum(x4046),
            total_bags = sum(total_bags))

# Missouth, Northeast, Plains, SouthCentral, Southeast, West, GreatLakes, California
# These are the regions that make up the total US. Above code shows that the total numbers of the categories are the same

# I will filter the data so that it only uses the lines for the above regions to avoid double counting.
# no need to worry about the smaller areas as these are contained within the bigger regions
# also i dont think it makes sense to look at regions that small
# also by the looks of it the remaining cities arent an exhaustive list 

# might be good to extract month from date as a categorical
# proportions of 4046, 4225, 4770 may be better than absolutes, also props of bag size may be better than absolute numbers 
```
```{r}
avocados %>% 
  filter(region %in% c("Midsouth", 
                       "Northeast", 
                       "Plains", 
                       "SouthCentral", 
                       "Southeast", 
                       "West", 
                       "GreatLakes", 
                       "California")) %>% 
  mutate(a = round(small_bags + large_bags + x_large_bags, 2) - round(total_bags, 2)) %>% 
  arrange(desc(a))

avocados



alias(lm(average_price ~ ., data = avocados))

# for some reason the bag sizes dont properly add up to the total_bags
# without the round they are different by like 0.000000000001
# with the round there are a couple at either end out by +-1 
# this means they dont show up in alias - I'm going to manual remove x larges bags as this can be derived from the other 3
# x4046, x4225 and x 4770 dont add up to the total volume so can keep them all
```

```{r}
avocados <- avocados %>% filter(region %in% c("Midsouth", 
                       "Northeast", 
                       "Plains", 
                       "SouthCentral", 
                       "Southeast", 
                       "West", 
                       "GreatLakes", 
                       "California")) %>% 
  mutate(month = as.factor(month(date)),
         across(where(is.character), as.factor),
         year = as.factor(year)) %>% 
  dplyr::select(-x1, -date, -x_large_bags) 
  
```

```{r}
avocados %>% 
  dplyr::select(average_price, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 

# total volume and total bags strongest

avocados_prop <- avocados %>% 
  mutate(across(small_bags:large_bags, ~ .x / total_bags))

avocados_prop %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 

#props don really help 




```

```{r}
n_data <- nrow(avocados_prop)

test_index <- sample(1:n_data, size = n_data * 0.2)

test <- avocados_prop %>% 
  slice(test_index)

train <- avocados_prop %>% 
  slice(-test_index)
```


```{r}
lm1 <- lm(average_price ~ total_volume, train)

summary(lm1)
plotModel(lm1)
plot(lm1)

# diagnostocs are mad i think this is because of the organic type



```

```{r}
avocados_resid <- train %>% 
  add_residuals(lm1) %>% 
  dplyr::select(-average_price, -total_volume) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(resid, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados_resid %>% 
 dplyr::select(is.numeric) %>% 
  ggpairs() 

```

```{r}
train %>% 
  ggplot(aes(x = x4046, y = average_price, colour = type)) +
  geom_point() +
  geom_smooth(method = "lm", se = FALSE)
```

```{r}
# filter by type to check if the are correlated with different things

avocados %>% 
  filter(type == "conventional") %>% 
  dplyr::select(average_price, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados %>% 
  filter(type == "conventional") %>%
  dplyr::select(is.numeric) %>% 
  ggpairs() 

# total volume and total bags strongest

avocados_prop %>% 
  filter(type == "conventional") %>%
  dplyr::select(is.numeric) %>% 
  ggpairs()




avocados %>% 
  filter(type == "organic") %>% 
  dplyr::select(average_price, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados %>% 
  filter(type == "organic") %>%
  dplyr::select(is.numeric) %>% 
  ggpairs() 

# total volume and total bags strongest

avocados_prop %>% 
  filter(type == "organic") %>%
  dplyr::select(is.numeric) %>% 
  ggpairs() 


#props don really help 
```



```{r}
# going to do type first

lm2 <- lm(average_price ~ type, train)

summary(lm2)
plot(lm2)


```


```{r}
avocados_resid <- train %>% 
  add_residuals(lm2) %>% 
  dplyr::select(-average_price, -type) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(resid, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados_resid %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 
```



```{r}
lm3 <- lm(average_price ~ type + region, train)

summary(lm3)
plot(lm3)

anova(lm2, lm3)

```


```{r}
avocados_resid <- train %>% 
  add_residuals(lm3) %>% 
  dplyr::select(-average_price, -type, -region) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(resid, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados_resid %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 
```


```{r}
lm4 <- lm(average_price ~ type + region + month, train)

summary(lm4)
plot(lm4)

anova(lm3, lm4)

# log helps diagnostics

lm4 <- lm(log(average_price) ~ type + region + month, train)

summary(lm4)
plot(lm4)

```

```{r}
avocados_resid <- train %>% 
  add_residuals(lm4) %>% 
  dplyr::select(-average_price, -type, -region, -month) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(resid, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados_resid %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 
```



```{r}

lm5 <- lm(log(average_price) ~ type + region + month + year, train)

summary(lm5)
plot(lm5)

anova(lm4, lm5)
```

```{r}
avocados_resid <- train %>% 
  add_residuals(lm5) %>% 
  dplyr::select(-average_price, -type, -region, -month, -year) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(resid, is.factor) %>% 
  ggpairs() 

# month, type and region all quite strong

avocados_resid %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 
```

```{r}
lm6 <- lm(log(average_price) ~ type + region + month + year + x4046, train)

summary(lm6)
plot(lm6)

```

```{r}
avocados_resid <- train %>% 
  add_residuals(lm6) %>% 
  dplyr::select(-average_price, -type, -region, -month, -year, -x4046) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 
```


```{r}



lm7 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume, train)

summary(lm7)
plot(lm7)
```

```{r}
avocados_resid <- train %>% 
  add_residuals(lm7) %>% 
  dplyr::select(-average_price, -type, -region, -month, -year, -x4046, -total_volume) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 
```

```{r}
lm8 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags, train)
lm8b <- lm(log(average_price)~1+type+year+region+month+total_volume+total_bags+small_bags+large_bags, train)

summary(lm8)
plot(lm8)
summary(lm8b)
plot(lm8b)
```

```{r}
avocados_resid <- train %>% 
  add_residuals(lm8) %>% 
  dplyr::select(-average_price, -type, -region, -month, -year, -x4046, -total_volume, -large_bags) %>% 
  dplyr::select(resid, everything()) 


avocados_resid %>% 
  dplyr::select(is.numeric) %>% 
  ggpairs() 
```

```{r}
lm9 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags, train)

summary(lm9)
plot(lm9)
```

```{r}
lm10 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume, train)

summary(lm10)
plot(lm10)
```



```{r}
lm11 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume + type:region, train)

lm12 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume + type:region + type:year + region:total_volume, train)

summary(lm11)
plot(lm11)
glance(lm11)
glance(lm12)
```


```{r}
train_rmse <- train %>% 
  add_residuals(lm12) %>% 
  mutate(sq_resid = resid^2) %>% 
  summarise(mse = mean(sq_resid),
            rmse = mse^0.5) %>% 
  pull(rmse)


train_rmse

predictions_test <- test %>% 
  add_predictions(lm12) %>% 
  add_residuals(lm12) %>% 
  dplyr::select(average_price, pred, resid) 
  
test_rsme <- predictions_test %>% 
  mutate(sq_resid = resid^2) %>% 
  summarise(mse = mean(sq_resid),
            rmse = mse^0.5) %>% 
  pull(rmse)

test_rsme / train_rmse
```


```{r}
cv_10_fold <- trainControl(
  method = "cv",
  number = 10,
  savePredictions = TRUE
)

model2 <- train(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume, 
               avocados_prop, 
               trControl = cv_10_fold, 
               method = "lm")

 model2$resample %>% 
  summarise(av_r2 = mean(Rsquared),
            av_rmse = mean(RMSE))
```

```{r}
cv_10_fold <- trainControl(
  method = "cv",
  number = 10,
  savePredictions = TRUE
)

model2 <- train(log(average_price) ~ type + region + month + year + x4046+ total_volume + large_bags + type:total_volume + type:region + type:year + region:total_volume, 
               avocados_prop, 
               trControl = cv_10_fold, 
               method = "lm")

 model2$resample %>% 
  summarise(av_r2 = mean(Rsquared),
            av_rmse = mean(RMSE))
```

```{r}
calc.relimp(lm9, type = "lmg", rela = TRUE)
```




AUTOMATED

```{r}

#### GARBAGE

regsubset_forwards <- regsubsets(average_price ~ ., 
                          data = avocados_prop,
                          nvmax = 40,
                          method = "forward")


summary(regsubset_forwards)

plot(regsubset_forwards,
     scale = "bic")

sum_forward <- summary(regsubset_forwards)

plot(sum_forward$bic,
     type = "b")
```

```{r}
# glmulti


glmulti_fit <- glmulti(
  log(average_price) ~ ., # model to fit, in this case, charges varies with everything
  level = 2, # level = 2 means try pairwise interactions. level = 1 means main effects only
  data = train, # data to use for fitting
  minsize = 0, # min size of model to try, in number of predictors
  maxsize = -1, # max size to try, set to -1 for unlimited
  marginality = TRUE, # marginality true means include pairwise interaction only if both main effects present in model.  
  method = "d", # method "d" means trial run, to get size of problem. Set to "h" for exhaustive search, or "g" for genetic algorithm
  confsetsize = 10, # how many models should glmulti() return? Must be less than total size of problem
  plotty = FALSE, # provide progress plots? Generally annoying.
  report = TRUE, # provide progress reports? Generally useful.
  fitfunction = lm, # use lm() as fit function. Can also use glm() for logistic regression.
  crit = aic # criterion for selecting best models. 
)
```


```{r}
lm_multi <- lm(log(average_price)~1+type+year+region+month+total_volume+x4046+x4225+x4770+total_bags+small_bags+large_bags+year:type+region:type+region:year+month:type+month:year+month:region+x4046:total_volume+x4225:total_volume+total_bags:x4225+total_bags:x4770+small_bags:total_volume+small_bags:x4046+small_bags:x4225+small_bags:x4770+small_bags:total_bags+large_bags:total_volume+large_bags:small_bags+type:total_volume+type:x4225+type:total_bags+year:total_volume+year:x4046+year:x4770+year:total_bags+year:large_bags+region:total_volume+region:x4225+region:total_bags+region:large_bags+month:total_volume+month:x4770+month:small_bags, train) 

summary(lm_multi)
glance(lm_multi)
```

```{r}
train_rmse <- train %>% 
  add_residuals(lm_multi) %>% 
  mutate(sq_resid = resid^2) %>% 
  summarise(mse = mean(sq_resid),
            rmse = mse^0.5) %>% 
  pull(rmse)


train_rmse

predictions_test <- test %>% 
  add_predictions(lm_multi) %>% 
  add_residuals(lm_multi) %>% 
  dplyr::select(average_price, pred, resid) 
  
test_rsme <- predictions_test %>% 
  mutate(sq_resid = resid^2) %>% 
  summarise(mse = mean(sq_resid),
            rmse = mse^0.5) %>% 
  pull(rmse)

test_rsme / train_rmse

# over fit? despite having a good bic
```
```{r}
glmulti_fit <- glmulti(
  log(average_price) ~ ., # model to fit, in this case, charges varies with everything
  level = 2, # level = 2 means try pairwise interactions. level = 1 means main effects only
  data = train, # data to use for fitting
  minsize = 0, # min size of model to try, in number of predictors
  maxsize = 10, # max size to try, set to -1 for unlimited
  marginality = TRUE, # marginality true means include pairwise interaction only if both main effects present in model.  
  method = "d", # method "d" means trial run, to get size of problem. Set to "h" for exhaustive search, or "g" for genetic algorithm
  confsetsize = 1000, # how many models should glmulti() return? Must be less than total size of problem
  plotty = FALSE, # provide progress plots? Generally annoying.
  report = TRUE, # provide progress reports? Generally useful.
  fitfunction = lm, # use lm() as fit function. Can also use glm() for logistic regression.
  crit = aic # criterion for selecting best models. 
)
```


```{r}
train = train %>% 
  select(average_price, region, type, year, month, total_volume, large_bags)


```





```{r}
lm_multi <- lm(log(average_price)~ 1 + region + type + year + month + total_volume + large_bags + year:type + month:year + region:total_volume + type:total_volume, train)

summary(lm_multi)
glance(lm_multi)
plot(lm_multi)

lm12 <- lm(log(average_price) ~ type + region + month + year + x4046 + total_volume + large_bags + type:total_volume + type:region + type:year + region:total_volume, train)
summary(lm12)
glance(lm12)
plot(lm12)
```

